DeepSeek Coder V2 Models by DeepSeek
Last refreshed 2026-10-17. Next refresh: weekly.
Details
About
DeepSeek Coder V2 is an open-source family of Mixture-of-Experts (MoE) code language models crafted specifically for code-related tasks. It builds on the advancements of the DeepSeek V2 model, featuring notable improvements in code-specific tasks and reasoning capabilities. These models were trained on an additional 6 trillion tokens, enhancing their skills in coding and mathematical reasoning while maintaining strong general language performance. Key enhancements include support for over 338 programming languages, a significant jump from the 86 supported in earlier iterations, and an increased context length of up to 128K tokens. The family offers a variety of models such as smaller "Lite" versions for projects with limited computational resources, and larger models for complex tasks. Available on Hugging Face, these models can be accessed through their API or chat interface, making them easily deployable across diverse coding environments 123.
Decision facts
- Best fit
- codingcodemath-heavy prompts
- Capability starting point
- DeepSeek Coder V2 Instruct (0724) with 128K context
- Lowest tracked input
- DeepSeek Coder V2 Lite Instruct · $0.120/1M · Novita AI
- Closest related family
- DeepSeek OCR
Archived Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work.
Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work.
Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work.
Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work.
Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| DeepSeek Coder V2 Instruct (0724) | Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work. | 2024-07 | code128K context236B parameters | Replaced |
| DeepSeek Coder V2 | Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work. | 2024-06 | code128K context236B parameters | Replaced |
| DeepSeek Coder V2 Instruct | Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work. | 2024-06 | code128K context236B parameters | Replaced |
| DeepSeek Coder V2 Lite | Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work. | 2024-06 | code128K context16B parameters | Replaced |
| DeepSeek Coder V2 Lite Instruct | Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work. | 2024-06 | code128K context16B parameters | Replaced |
Release Timeline
2 release groupsReplaced By
Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work.
Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work.
Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work.
Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work.
Keep for legacy integrations; evaluate DeepSeek V4 Pro before new work.
Specifications(5 models)
| Model | Released | Context | Parameters |
|---|---|---|---|
| DeepSeek Coder V2 Instruct (0724) | 2024-07 | 128K | 236B |
| DeepSeek Coder V2 | 2024-06 | 128K | 236B |
| DeepSeek Coder V2 Instruct | 2024-06 | 128K | 236B |
| DeepSeek Coder V2 Lite | 2024-06 | 128K | 16B |
| DeepSeek Coder V2 Lite Instruct | 2024-06 | 128K | 16B |
Available From(3 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| DeepSeek Coder V2 Lite Instruct | Novita AI | $0.12 | $0.36 | Serverless |
| DeepSeek Coder V2 | DeepSeek Platform | $0.14 | $0.28 | Serverless |
| DeepSeek Coder V2 Lite | Fireworks AI | $0.5 | $0.5 | Serverless |
| DeepSeek Coder V2 Instruct | Fireworks AI | $1.2 | $1.2 | Serverless |
Popular comparisons in this family
Related Model Families
Models(5)
DeepSeek Coder V2 Instruct (0724)
DeepSeek Coder V2
DeepSeek Coder V2 Instruct
DeepSeek Coder V2 Lite
DeepSeek Coder V2 Lite Instruct